How to get drag coefficient for a rocket?

[martincmartin]

I'm looking to model KSP's dynamics to do some offline optimization.  Modeling gravity and thrust are pretty easy, but modeling drag is harder.  I can get the density at all points along my flight path, but to turn that into the drag force, I need to know the overall drag coefficient times the reference area.  Computing it from the drag cubes of all parts seems like a lot of work that's hard to get right, and the engine does it for us anyway.  Can we get access to it?  I can't find anything relevant in the Vessel class.  My fallback would be to take the total acceleration, subtract out gravity and thrust, divide by density and speed squared, and pray that the answer is both accurate and numerically stable.

[martincmartin]

The AeroGUI mod does this, it loops over each part, summing the drag.

[NathanKell]

Drag coefficient is meaningless without also getting reference area. Aero GUI will give you the product (i.e. the Cd * S combined term in the drag equation) but in order to compute Cd you need to decide on how to compute S (and then you just divide the Cd*S term by S, to get Cd).

[wasml]

I may be doing something wrong but I found the atmospheric density from GetPressure(alt), GetTemperature(alt), GetDensity(P, T) differed from part.GetAtmDensity() by a significant bit.

The first method read 1.115  while the second read 1.214 while sitting on the runway.

Might throw your calculations off if your using the first method.

(or do you know a better way to get atmospheric density at a remote location?)

 

[NathanKell]

GetTemperature() gets the reference temperature at that altitude. It does not get the current temperature; that depends on your location (latitude) and the time of day.

[martincmartin]

Thanks NathanKell, Cd * S is really what I want.  I would have thought it was a static property of my rocket, but even for a very simple rocket traveling straight up it changes quite a bit, from 0.675 at launch to 0.24 then stabilizes at 0.30.  I thought Cd * S would be independent of atmospheric density, ship velocity, altitude and everything else, and depend only on ship geometry and wind direction, which is constant in this case.  Any ideas?

Edited March 1, 2016 by martincmartin

[NathanKell]

Cd is dependent on Mach, and on Reynolds number (or, in stock KSP, pseudo-Reynolds).